Delta DPS-280FP suddenly not starting after recap

[Behemot]

I have one Lenovo ThinkCentre A55 tower computer here and I decided to make some preventive recap of this PSU because Teapo's on motherboard were already bulged. MoBo works like charm after complete recap.

I have been doing testing of the PSU during recap becuase I learned so far that its difficult to find problems when it is all done and suddenly does not work Yet the PSU fired up after each few replaced caps all the way to complete recap. That was yesterday. But today, after assembling the whole system it just blinks and shuts down.

When trying to start just the PSU alone, fan spins for half a second, but it turns off immediatelly and orange diode blinks. +5V SB is just fine and I havent found any short on the output. So what may be the problem? I did some messy soldering here and there but all the joints seem and measure fine.

May it be with the caps used again? All the big ones were L'Tecs except for one Nicichon but i swapped them for Samxon's 3300 uF in 10 and 16 V values and one Chemi-Con KY (capacitance on +5V SB was increased). Small L'Tec and Taicon ones changed for: one Chemi-Com KZG, Rubycons CE (W) from old telly and some other two greater ones (some Rubycon and Nichicon from old CRT display).

[tj2]

I've found that in some PSUs there is a cap, perhaps a 22 µF 50 V, that will prevent start up if it has an ESR that's too different from the original. I recapped a PSU recently that wouldn't start up afterwards. The 5 VSB was fine but when I shorted the power-on line, the PSU would briefly spin the fan and light up my dummy load, but then immediately shut down. When I replaced the new Panasonic FM 22 µF 50 V cap by a Nichicon TT, it worked correctly. The Nichicon was reasonably close to the original cap in ESR but the Panasonic had a fairly low ESR, about a third of the original.

[Behemot]

Only problem is I dont have any idea what ESR the Rubycons CE W have…they are so old that there is no datasheet. But they are fine, when I masured, capacitance was at some 90 % of nominal and ESR lowest which the meter could show (some diode thing).

Yeah, seems that the original LTec has ripple of only 66 mA What if I changed it for two 10uF Chemi-Cons KMGs in paralel?

ADD// swaped it for the original one, no change…gonna try also the other small ones…

ADD2// bingo! 10 uF LTec I love U man…write down one beer on me

[tj2]

If you want to replace the 10 µF LTec, you might try a Nichicon PW, Panasonic FC, or something similar. Probably a general purpose cap will do as long as it's 105 C rated. What voltage and series is the LTec?

[Behemot]

I changed it for Chemi-Con KMG, thats universal series. I also have the Nichicons PW, but these are not GP IMO, they have way better specs than this LTec crap. It was also TK series, 50 V.

[tj2]

As long as the KMG works, I'd leave it in. Its ripple rating is 40 mA rather than 66 mA, but that probably doesn't matter. Perhaps someone else has a better idea of how stressed this cap is. I'd guess not very.

[Behemot]

It sits on the compansation, right next to one ov the driving ICs.

Well, seems that I have to use the same cure for one Fortron which whuts down in not a nice maner withing 5 seconds. Where are the old times when these PSU's handled all the nice caps you soldered in instead of the craps? Why the stupids design the driving circuitry for bad caps now?